Renal dysfunction following cardiac surgery is more apparent in high-risk patients with pre-existing renal dysfunction, diabetes and impaired left-ventricular function, and following complicated procedures involving prolonged cardiopulmonary bypass (CPB). The aim of this prospectively randomised double-blinded placebo-controlled study was to evaluate reno-protective effect of low-dose furosemide infusion in this high-risk group. Patients with preoperative serum creatinine >130 micromol/l (1.4 mg/dl), left-ventricular ejection fraction <50%, congestive heart failure, diabetes, or procedures involving prolonged CPB were randomised to receive either saline at 2 ml/h (n=21), or furosemide at 4 mg/h (n=21). Infusion was commenced after induction of anaesthesia and continued for 12h postoperatively. Renal dysfunction was defined as >50% increase in serum creatinine postoperatively, or >130 micromol/l (1.4 mg/dl), or requirement for haemodialysis, or all of these. In patients with preoperative serum creatinine >130 micromol/l, >50% increase over preoperative levels was used to define postoperative renal dysfunction. Following cardiac surgery, patients receiving furosemide had a higher urine output (3.4+/-1.2 ml/kg/h in furosemide group and 1.2+/-0.5 ml/kg/h in placebo group; p<0.001), higher postoperative fluid requirement (4631+/-1359 ml in furosemide group and 3714+/-807 ml in placebo group, p=0.011), and lower urinary-creatinine (2+/-1.3 micromol/l in furosemide group and 5.9+/-2.5 micromol/l in placebo group p<0.001). Both groups had significant increase in retinol binding protein/creatinine ratio (7.2+/-6 to 3152+/-1411 in furosemide group; 4.9+/-2.1 to 2809+/-1125 in placebo group; p<0.001) and peak serum creatinine (98+/-33 to 177+/-123 micromol/l in furosemide group; 96+/-20 to 143+/-87 micromol/l in placebo group; p<0.001), and a significant decrease in peak creatinine-clearance (64.3+/-29.4 to 39.1+/-16.6 ml/min in furosemide group; 65.5+/-38.6 to 41.8+/-17.8 ml/min in placebo group; p<0.001) following cardiac surgery, implying significant renal injury following cardiac surgery. Peak creatinine levels (177+/-123 micromol/l in furosemide group and 143+/-87 micromol/l in placebo group; p=0.35) and peak creatinine-clearance (39.1+/-16.6 ml/min in furosemide group and 41.8+/-17.8 ml/min in placebo group; p=0.61) were similar in the two groups. Importantly, there was no difference in incidence of renal dysfunction between the furosemide group (9/21) and the control group (8/21) (relative risk 1.1, 95% confidence interval 0.6-2.2; p=0.99). Our randomised trial did not demonstrate any benefit of furosemide-infusion postoperatively in high-risk cardiac surgical patients. Although urinary output increased with furosemide, there was no decrease in renal injury, and no decrease in incidence of renal dysfunction.